A new era of chips is about to begin.

Recently, in the article "Chips, the Latest Roadmap", the author shared IMEC's predictions for the semiconductor roadmap over the next 14 years.

IMEC's predicted roadmap up to 2039

It can be seen that with the evolution of advanced process nodes and the innovation of transistor architectures, two-dimensional semiconductor materials may become the focus of the industry in the future.

In fact, currently, Moore's Law is slowing down. As the process nodes approach the physical limit, the manufacturing structure of silicon-based three-dimensional transistors has become increasingly complex, the cost required has increased exponentially, while the marginal benefits brought by technological evolution have significantly decreased.

Meanwhile, to maintain the progress of Moore's Law, the focus of innovation has shifted from size scaling to functional scaling. In the arduous journey from FinFET to Nanosheet and even future CFET and other transistor architectures, the industry has deeply realized that relying solely on the three-dimensional stacking technology of silicon-based materials is no longer sufficient to support sustainable miniaturization and energy efficiency improvement. Seeking fundamental material innovation has become the key to breaking through the bottleneck and opening up a new growth curve.

Against this background, the strategic transition from traditional silicon-based three-dimensional materials to two-dimensional semiconductor materials has rapidly become the core focus of global semiconductor R & D and industrial layout, attracting the attention of global researchers and the industry.

Two-dimensional Semiconductors: A Formidable Force

As is well known, as semiconductor processes approach the sub-nanometer level, silicon-based devices encounter physical limits such as thickness fluctuation scattering, quantum tunneling effect, and short-channel effect, leading to significant performance degradation, which has become the main obstacle to the continuation of Moore's Law. Although three-dimensional stacking technology can continue the growth of transistor density, using traditional channel materials for 3D integration will be extremely challenging, and the increased dependence on the nanoscale alignment accuracy of EUV lithography has exacerbated the cost pressure.

Therefore, the introduction of two-dimensional materials as channel materials provides an innovative solution to the challenge of size miniaturization.

With atomic-level thickness (0.3 - 10nm) and van der Waals heterojunction technology, two-dimensional materials can construct vertical field-effect transistors (VFETs) to achieve a density breakthrough 10 times that of FinFETs, and still maintain a switching ratio of 10⁶ at a gate length of 1nm. Their unique electrical properties (such as the mobility of black phosphorus up to 60000 cm²V⁻¹s⁻¹) and quantum properties (such as the superconducting state of magic-angle graphene and the valley polarization effect of tungsten diselenide) make them ideal channel materials for next-generation integrated circuit chips.

Meanwhile, two-dimensional materials have advantages for chip applications. Different from traditional bulk silicon materials, two-dimensional materials exhibit lattice periodicity in the plane. By controlling the geometric structures such as the number of layers and heterostructures of two-dimensional materials, or applying external strain and electric fields, their lattice periodicity can be changed, ultimately affecting the energy band structure and the size of the bandgap. The different lattice structures and atomic arrangements of two-dimensional materials result in different electronic energy band structures, thus producing a wide energy band range, covering semi-metals, semiconductors, and insulators. Two-dimensional materials are easy to integrate with other materials without being restricted by lattice constant matching.

With the advent of the AI and big data era, some new computing architectures and mechanisms have been introduced into next-generation computing technologies. Benefiting from the unique properties of two-dimensional materials, storage devices, neuromorphic devices, quantum devices, ion transistors, etc. based on two-dimensional materials have been widely studied and applied, becoming the core engine to break through the physical limit in the post-Moore era.

The most typical and earliest experimentally proven two-dimensional material is graphene. In 2004, K. S. Novoselov et al. published an article in the journal Science, reporting that graphene was obtained from highly oriented pyrolytic graphite by mechanical exfoliation, and its unique and excellent electrical properties were demonstrated.

As the first discovered two-dimensional material, graphene with a thickness of only 0.335 nanometers is considered the most promising alternative material for semiconductors. It has extremely excellent physical properties, such as high strength, high conductivity, and high thermal conductivity. The scientific community hopes to use it to prepare a new generation of semiconductor devices, and it is a strong candidate material for the next-generation "carbon-based semiconductors".

Previously, a study by IBM showed that compared with silicon-based chips, graphene chips are expected to have significant improvements in performance and power consumption. For example, a 7-nanometer process graphene chip is expected to be up to 300% faster than a 7-nanometer process silicon-based chip - provided that a "gap" can be opened in the energy band of graphene.

However, as a semi-metallic material, the zero-bandgap characteristic of graphene makes it impossible to achieve the ideal state of semiconductor current shutdown, and it is difficult to be made into electronic switching elements like transistors, which limits its application in logic devices. Therefore, although the K. S. Novoselov team prepared graphene, they still showed pessimism in a review article about graphene in 2007.

For a long time after the birth of graphene, it really failed to find a "use" in the semiconductor field. Nevertheless, two-dimensional materials represented by graphene have still received great attention, and new two-dimensional materials have emerged like mushrooms after rain.

Since the successful isolation of graphene in 2004, two-dimensional material systems such as transition metal dichalcogenides (TMDCs, such as MoS₂ and WS₂), hexagonal boron nitride (h-BN), black phosphorus, and MXene have been widely studied. Especially after 2010, the successful preparation of single-layer MoS₂ transistors marked the entry of two-dimensional semiconductors into the practical stage.

In 2024, a research team composed of researchers from Tianjin University in China and the Georgia Institute of Technology in the United States made a major breakthrough by using a special furnace to grow epitaxial semiconductor graphene monolayers on silicon carbide wafers. The research found that if properly manufactured, epitaxial graphene will chemically bond with silicon carbide and exhibit semiconductor characteristics, which successfully overcomes the key technical problem that has long hindered the development of graphene electronics, opens the energy bandgap of graphene, and achieves a breakthrough from "0" to "1". The relevant paper was published in Nature. Therefore, graphene has also had a "new life".

Thanks to many advantages, since the discovery of graphene, two-dimensional materials have gradually become a large family with many members and diverse categories. These common two-dimensional semiconductor materials each have different energy band structures and electronic characteristics, covering material types from superconductors, metals, semi-metals, semiconductors to insulators, and also have excellent optical, mechanical, thermal, magnetic, and other properties.

Crystal structures and properties of some typical two-dimensional materials

• Graphene: A Zero-Bandgap Dirac Fermion System. The electronic structure of graphene shows linear dispersion, forming a Dirac cone at the K point. The carrier mobility is extremely high, reaching 10⁴ - 10⁵ cm²/V·s at room temperature. Its quasiparticle behavior is approximately massless, and the spin divergence length is long, making it suitable for high-frequency electronics and spin transport research. However, the zero-bandgap limits its application in digital switching devices, and bandgap engineering methods are needed to achieve energy band regulation.
• TMDs: Direct Bandgap and Valley - Spin Coupling. Typical TMDs (such as MoS₂ and WS₂) exhibit direct bandgap semiconductor properties in the single - layer state, with a bandgap of about 1.8eV, accompanied by strong spin - orbit coupling and broken spatial inversion symmetry, resulting in the coupling of spin and valley degrees of freedom. This physical mechanism makes TMD an ideal platform for valleytronics and optical spin manipulation. Although its mobility is relatively low, its stable bandgap and excellent photoelectric response make it have practical application value in transistors and photodetection.
• Black Phosphorus: Anisotropy and Tunable Bandgap. Black phosphorus is a rare intrinsic direct bandgap material. Its energy band structure is sensitive to the number of layers, and can be continuously tuned from 2eV for a single layer to about 0.3eV for a bulk. At the same time, the lattice structure of black phosphorus leads to strong electronic anisotropy, giving it unique application prospects in direction - related devices. Although its mobility can reach the order of thousands of cm²/V·s, its chemical instability is a major obstacle restricting its development.
• MXene: A Two - Dimensional Metal and Interface Engineering Platform. MXene is a class of layered transition metal carbides/nitrides. Its natural metallicity, high electrical conductivity, and surface functional groups endow it with broad potential in contact engineering, electrochemistry, and tunable energy band design. Research shows that by adjusting the surface terminal groups and interface stress, it can be induced to transform from a metallic state to a semiconductor state, making it one of the most engineerable systems among two - dimensional materials.

Benefiting from the quantum confinement effect in the atomic layer thickness direction, these two - dimensional materials exhibit properties that are completely different from their corresponding three - dimensional structures. The transition from silicon - based three - dimensional materials to two - dimensional semiconductor materials is not only a material innovation but also a leap in semiconductor technology, which is expected to break the deadlock of the slowdown of Moore's Law and push the semiconductor industry into a new stage of development.

Professor Ren Tianling and Associate Professor Tian He from Tsinghua University, Professor Liu Ziyu from Fudan University, Professor Guo Hao from North University of China, Associate Researcher Peng Song'ang from the Institute of Microelectronics of the Chinese Academy of Sciences, and Professor Deng Tao from Beijing Jiaotong University, etc. summarized the research progress of two - dimensional semiconductors in process engineering and various chip application fields.

Schematic diagram of the general roadmap for two - dimensional circuits. (a) The development timeline of silicon - based, carbon nanotube - based, and two - dimensional - based integrated circuits. (b) The implementation route of two - dimensional circuits and possible future application fields.

Large enterprises such as TSMC, Intel, Samsung, and IMEC have accelerated their layout in the two - dimensional semiconductor track, investing a large amount of capital in the research and integration of two - dimensional semiconductor materials, and promoting the industry to move from the laboratory to large - scale production.

Data shows that in 2024, the global market size of two - dimensional semiconductor materials reached 1.8 billion US dollars, among which graphene was the largest segment market, accounting for 45% mainly due to its superior conductivity and mechanical strength. Transition metal dichalcogenides (TMDs) became the second - largest segment market due to their unique electronic properties and versatility in various applications, accounting for 30%. With the maturity of preparation technologies, it is expected that the market size will expand at a compound annual growth rate of 24% - 26.5% from 2025 to 2030, and is expected to exceed 4.5 billion US dollars in 2030. The main growth drivers come from the demand in the fields of 5G communication, AIoT, and high - performance computing.

Under this trend, research institutions and the industry have actively explored two - dimensional semiconductor materials and devices, promoting the research and development of two - dimensional semiconductor materials.

Progress and Breakthroughs in the Two - Dimensional Semiconductor Industry

Yuanjiwei: The First Domestic Engineering Demonstration Line for Two - Dimensional Semiconductors Launched

In June 2025, the engineering verification and demonstration process line for two - dimensional semiconductors of Yuanjiwei Technology, incubated by a research team from Fudan University, was launched in Chuansha, Pudong. This is the first domestic engineering demonstration line for two - dimensional semiconductor integrated circuits. Yuanjiwei plans to build a commercial mass - production line within three years, tackle core technologies such as front - and back - end processes, compatibility between "non - silicon" materials and silicon - based processes, and heterogeneous/heterogeneous integration. It has cooperated with enterprises such as Zhongke Chuangxing and Beijing Saiwei Electronics to explore heterogeneous integration solutions.

Relying on the advantage of extremely low leakage current of two - dimensional materials, Yuanjiwei has selected DRAM and edge computing as the entry points for industrialization, and is promoting the integrated development of prototype chiplets. It will complete process optimization on an 8 - inch engineering demonstration line. In April this year, its joint team published the world's first 32 - bit RISC - V architecture microprocessor "WUJI" based on two - dimensional semiconductors in Nature. It integrates 5,900 transistors, with an inverter yield of 99.77%, refreshing the integration record and improving the performance by 51 times. This processor is based on single - layer molybdenum disulfide (MoS₂), does not rely on EUV lithography machines, and realizes full - chain independent R & D, proving the feasibility of applying new materials.

Yuanjiwei aims to become the "TSMC" in the field of two - dimensional semiconductors. It plans to break through the compatibility problem between materials and silicon - based processes within three years, build an internationally leading demonstration commercial production line, and achieve the performance of 1 - 2 nanometer chips.

Northwest Institute for Non - Ferrous Metal Research and Xi'an Institute for Rare Metal Materials: Synthesis of Ti₂CO₂ Two - Dimensional Material by POT Technology

Biosensors rely on inorganic/organic composite structures, and the affinity layer is the connecting core. MXene materials have become potential candidates due to their unique physical and chemical properties. Among them, Ti₂CO₂ MXene has attracted attention due to its high stability and semiconductor characteristics. However, conventional preparation methods are prone to over - oxidation to form TiO₂, making it difficult to synthesize accurately.

The teams of Li Yang and Cheng Fei from the Northwest Institute for Non - Ferrous Metal Research and the Xi'an Institute for Rare Metal Materials developed the ozone pulse treatment (POT) technology and successfully synthesized stable two - dimensional semiconductor Ti₂CO₂ MXene. This technology precisely controls the activation energy window of the reaction "Ti₂C→Ti₂CO₂→TiO₂", uses highly active ozone to lower the energy barrier of the target reaction, and suppresses over - oxidation through short - time pulses, breaking through the bottleneck of traditional functional group regulation. The high adsorption capacity and stability of Ti₂CO₂ MXene make it an excellent affinity layer for high - precision biosensing, providing new ideas for the functionalization and stabilization of MXene materials and expanding the application of two - dimensional semiconductors in the fields of biosensing, health monitoring, and intelligent diagnosis and treatment.

Nanjing University of Aeronautics and Astronautics/Nanyang Technological University: Coherent Confined Single - Metal Atomic Chains in Two - Dimensional Semiconductors

Recently, the teams led by Academician Guo Wanlin, Professor Zhang Zhuhua, and Associate Researcher Qiao Ruixi from Nanjing University of Aeronaut